Calibration of urban parameter (URBPARM.TBL) and anthropogenic heat

[shuchang769]

Hello, wrf team
I am recently doing a comparison of the simulations with single urban class and multiple urban class (low-density residential, high-density residential and industrial/commercial) based on the MODIS 21 types land cover scheme.
While I have realized that before and after using the multiple urban class, the simulation results did not change much.
So I am looking into the URBPARM.TBL file to search for the reason.
The URBPARM.TBL file I am using is the default one from WRF.
and the version of WRF I am using is wrf/3.9.1.1.
My question is:

• Is there any recommended method to calibrate the parameters listed in the URBPARM.TBL. Because there are too many parameters and they could be quite different for different cities. Is there any recommended data source for checking these parameters for different cities? Such as the proportions of the buildings of different heights, the thermal properties(heat capacity, thermal conductivity, albedo, emissivity, roughness length .etc) for different urban surfaces?

• For the anthropogenic heat configurations, I have noticed that in the default URBPARM.TBL the AH and ALH have been assigned with some default values. While the AHOPTION and ALHOPTION are both set to 0, which means no anthropogenic heating and no anthropogenic latent heat. Should I change them to 1 to activate the anthropogenic heating effects which has been defined in AH and ALH?

I would be grateful if anyone can help me and answer these questions!

Thanks!

Chang

 

[Ming Chen]

Chang,

The parameters for urban land surface are highly case-dependent. The default values given in WRF are based on a few observations and I don't think they are representative. One must adjust these values based on one's specific case. However, we don't have more information about how to specify the values. there is little publications discussing this issue. So they are more empirical.

For your second question, yes you need to set AHOPTION and ALHOPTION to 1 if you want to activate anthropogenic heating. Again, note that these values are also case-dependent. They may not well apply to your case. You probably need to do lots of tuning.

 

[Chang_Shu]

Thanks, Ming for the reply.
I have noticed that anthropogenic heat is only available for the UCM model (sf_urban_physics=1), but I have just run a case with BEP model(sf_urban_physics=2) with AHOPTION and ALHOPTION =1 successfully.
So I am wondering if the anthropogenic heat simulation is also compatible with BEP in WRF/3.9.1.1.

From my validation of the WRF model with multiple weather stations in the urban area, it was found that

• the simulated 2-m air temperature has a mean bias error of around -1.0.

• while the 10-m wind speed has a mean bias error of around +2.4.

So I am thinking of these 3 approaches below to calibrate the model:

• increase the lower boundary conditions of urban surface temperatures,

• reduce the albedo of urban surfaces

• and increase the roughness length to reduce the wind speed

Would you please confirm if these directions are correct to explore for the calibration of the model?
And please suggest a little bit if there are any other parameters that can be adjusted to optimize the results on wind speed and air temperature?


Thank you!

Chang

 

[Ming Chen]

Chang,

The anthropogenic heat is only available for the UCM model (sf_urban_physics=1). If you run with BEP/BEM model, even if you turn on the anthropogenic heat option, it has no impact on the model results. BEP/BEM is not aware of this option.

I agree with your ideas how to tune the model to obtain better results. Please try and let me know how they work.

Thanks.

 

[Chang_Shu]

Hello, Ming
I have tested one case by increasing the lower boundary temperature for building surfaces.
And the RMSE and MBE of T2 and precipitation come to be much better.
But the wind speed is still not that good (MBE~+2.4 at multiple weather stations).

So I am considering making changes to the roughness and building height-related settings.
I have noticed that in URBPARM.TBL:

• 1. there is roughness length for momentum over building walls(Z0B), ground(Z0G) and roof(Z0R). But Z0B and Z0G are only active for CH_SCHEME==1. but for CH_SCHEME, it was marked that the option is only applicable for sf_urban_physics=1. Does that mean I can only change Z0R since I am using sf_urban_physics=2?

• 2. most of the building shape and urban morphology settings are for sf_urban_physics 1 and 3, even for the roof level (building height) ZR is also only applicable for sf_urban_physics=1. Then for my case sf_urban_physics=2, it won't take effect if I change ZR.

• 3. So I can only change the percentage table for street and building height in the URBPARM.TBL to adjust the building shapes. Please help confirm if my understandings are correct.

To modify the roughness to reduce the wind speed in the simulation.
I have also noticed there are 3 files have roughness related information:

• 1. Z0R in the URBPARM.TBL

• 2. Z0MIN, Z0MAX and Z0 in the VEGPARM.TBL

• 3. SFZ0 in the LANDUSE.TBL.

I am wondering what is the difference between these parameters in different TBL files and which is the parameter I can change to calibrate the wind speed variable?

Also, please let me know if there are any other parameters that I can try for calibration of the wind speed.

Thanks,

Chang

 

[Ming Chen]

Please see my answers below:

1. there is roughness length for momentum over building walls(Z0B), ground(Z0G) and roof(Z0R). But Z0B and Z0G are only active for CH_SCHEME==1. but for CH_SCHEME, it was marked that the option is only applicable for sf_urban_physics=1. Does that mean I can only change Z0R since I am using sf_urban_physics=2?

Yes correct.

2. most of the building shape and urban morphology settings are for sf_urban_physics 1 and 3, even for the roof level (building height) ZR is also only applicable for sf_urban_physics=1. Then for my case sf_urban_physics=2, it won't take effect if I change ZR.

Yes you are right.

3. So I can only change the percentage table for street and building height in the URBPARM.TBL to adjust the building shapes. Please help confirm if my understandings are correct.

Yes this is the way to change the building shape.

To modify the roughness to reduce the wind speed in the simulation.
I have also noticed there are 3 files have roughness related information:
1. Z0R in the URBPARM.TBL
2. Z0MIN, Z0MAX and Z0 in the VEGPARM.TBL
3. SFZ0 in the LANDUSE.TBL.
I am wondering what is the difference between these parameters in different TBL files and which is the parameter I can change to calibrate the wind speed variable?

When you turn on urban scheme, Z0R in the URBPARM.TBL will be used over those points defined as urban landuse type. For all other points, data from
VEGPARM.TBL will be used. If you don't run with urban scheme, then data from URBPARM.TBL will not bee used. For all landuse type, the values from the VEGPARM.TBL will be used. LANDUSE.TBL is not used when running with Noah.

Note that even for the same urban landuse type, the values in the above three TBL files are different. This is because, as I mentioned previously, these values are empirical and based on a few on-site observations. Their values may evolve with time. Eve in different versions of WRF, their values can be different.

Note that there exist feedback loops in the physics processes. For example, temperature change may affect stability, which in turn influences mixing and winds. It is hard to tell what tuning can make the model perform as expected.

 

[Chang_Shu]

Thank you very much for the detailed answer.
So in my case, to reduce the wind speed prediction, I will try to only increase the Z0R in URBPARM.TBL and maybe also change the building height and street width percentage table to see if it can make the results better.

Chang

 

[Ming Chen]

Chang,

This is the approach I would follow if I want to do the same thing.